David sent me an email with lots of small tasks for this project. I've put them up on my brand new to-do list.

Ok, so here's how this all works: every point where |y| &le 0 that isn't occulted by the star gets numerically integrated. Points that are occulted by the star get the value 10¹⁰. Everywhere else uses the definite integral. And that's that. I've put up a bunch of images (most of them of the work in progress) on my images page.

Ok, so for the past couple of days I've been thinking about this whole "integrating &tau" thing, and it's been really frustrating me. I managed to get the indefinite integral for p &le 1, but it involves an inverse hypberolic tangent. IDL doesn't have a function for that, so I found a useful definition online. The only problem is that I can't use it for all regions of the data. So I was back to square 1.

Next I found qsimp, an IDL numerical integration routine. While reading about it online, I discovered that it sucks (big surprise), but that there's a better version called newsimp. These functions take the function over which one wishes to integrate as a variable, which poses a serious problem for me. I need to integrate over x for different values of y, which really isn't integrating over 2 variables. So, I needed to find a way to pass in y to the function. This is impossible to do directly while using newsimp (or for that matter qsimp). So I got clever. I found out how to declare global variables in IDL, which is surprisingly not straightforward (check out the ol' IDL page for more). But it turns out that you can't loop over global variables (check out the ol' IDL sucks page for more), but I managed to circumvent that, so I think I'm on my way to solving this problem, although it still remains to be seen if newsimp can do everything that I need it to in terms of regions in the data that are otherwise difficult, but we'll see.

First thing I did was set up the new computer, which I've (temporarily) named Wilkinson in honor of our fallen, inflatable comrade. Stan emailed me back, and I not only have the equation for &tau for p &le 1, but the code he gave me pointed out that I should take occultation into account at some point. I can do this either by setting &tau to a very large number in the occulted regions, or I can just skip those regions it in the line profile determination.

I poked around with Stan's code for a while before becoming confused and a bit frustrated. I can't figure out a general form of his equation...

I had a talk with David today about calculating &tau and the line profiles before he left for Germany. I've got &tau pretty much down, except for where |p| &le 1. For p=1, the relevant equation is undefined. I've posted the paper that discusses this, as well as discussed calculation difficulties on my IDL page. In addition, I've posted some new images. I also emailed Stan for some assistance.

I got a couple of things done today after watching Italy beat Ghana 2-0. First, I worked around a bug in my code that was causing errors when trying to save the state of the GUI to the 'lastgood.set' file. I still don't know what the error is, but oh well for now. I also managed to get the density scaling of the color brightness in the contour plots working after some clever deeds. Right now I'm stuck in that I don't know how to perform the integral on the equation for &tau_*, so I can't yet try to draw those contours on top.

I began an image naming system that's a little blunt but works. Basically, the filename consists of the four variable names (R_o, beta, q, and tau) and their values. I really can't think of any other way to differentiate between two different images made by this program, other than manual naming, which would be a huge pain in the ass I think.

I finally read the first article on this topic, and it's been posted on my articles page. This looks like it should be a fairly straightforward project, although we'll just have to wait and see.

While waiting to talk to David about this project, I created the first version of the GUI for this program. It's up on my images page.

After getting pretty frustrated with the progress being made on my post-processing project, David introduced me to this new project. Basically, I need to make analytic wind models and create visuals from them, including (of course) contour plots and line profiles.